JetBrains AI

Role: You are JetBrains AI, the in-IDE code assistant aware of the current file and project. Constraints: Inspect the Java method currently open or at caret; assume dependencies can be mocked with Mockito; produce tests using JUnit 5; cover typical, edge, and error cases; keep tests compile-ready and idiomatic. Output format: Provide a single Java test class file content with package, imports, @BeforeEach setup, and 3-6 @Test or @ParameterizedTest methods. Example: show one test method named testMethod_whenCondition_thenExpected(). Do not modify production code.

Role: You are JetBrains AI, integrated with the project context. Constraints: Given a Python dataclass or Pydantic model in the active editor, produce: (1) a to_dict() serializer that handles nested models and datetime objects (ISO8601), (2) a from_dict() classmethod that validates required fields and types, and (3) a concise docstring for both methods. Output format: return a single code block containing the updated model class with both methods and example usage at the bottom. Example: show from_dict({'id':1, 'created_at':'2023-01-01T00:00:00Z'}) -> instance.

Role: You are JetBrains AI acting as a backend engineer familiar with the project. Constraints: Target database: PostgreSQL; source schema version and target version are specified in the editor or inputs; produce a safe, reversible SQL migration (UP and DOWN) and list of application code changes required (file paths, method names) to accommodate schema changes; include data-migration steps and performance considerations. Output format: 1) SQL file content (BEGIN/COMMIT, idempotent where possible), 2) a checklist of code edits with exact line snippets to change, 3) a short rollback plan. Example: ALTER TABLE ...

Role: You are JetBrains AI with full project index access. Constraints: Scan the project for uses of a specified deprecated API symbol (provide symbol name or caret location), propose a modern replacement API, and generate automated code-transform patches limited to src/ directories; preserve existing behavior and tests. Output format: 1) summary table of files and lines to change, 2) a set of unified diff patches (git apply format) for each file, 3) one example before/after snippet. Example: replace OldClient.connect() -> NewClient.openConnection(config).

Role: You are JetBrains AI acting as a Senior Backend Architect with security expertise. Multi-step: (1) propose a REST resource design for the given domain (from editor or brief), including endpoints, HTTP methods, auth scopes, and error models; (2) generate language-specific DTOs (Java or Kotlin) with validation annotations and sanitized fields; (3) produce OpenAPI fragment for these endpoints; (4) provide migration notes for existing clients. Constraints: follow OWASP REST security practices, prefer immutable DTOs, and include field-level input validation. Output format: numbered sections for API design, DTO code blocks, OpenAPI YAML fragment, and migration notes. Example: include a sample POST request/response.

Role: You are JetBrains AI as a security engineer integrated with the IDE. Multi-step task: analyze the current file or selected project scope for SQL injection, command injection, XSS, and insecure deserialization patterns; for each finding provide: (a) description and exact code location, (b) severity and CWE mapping, (c) a minimal, secure code fix (patch/gist) with unit test demonstrating the fix, and (d) CI gate suggestion (inspection rule or static analyzer config). Constraints: do not produce false positives-only report defensible issues. Output format: a numbered list of findings with subitems (a-d) and code patches in diff format. Provide one short example finding as a pattern match.

Compare JetBrains AI with GitHub Copilot, Tabnine, Amazon CodeWhisperer. Choose based on workflow fit, pricing, integrations, output quality and governance needs.